Assessing Interviewer Effects in a Longitudinal Study of Bereaved Elderly Adults

Abstract

In a 2-year study, bereaved elders who were personally interviewed were compared with those who completed a mailed questionnaire on measures of depression, life satisfaction, and five global measures constructed from a series of bereavement-related feelings and behaviors. Data were gathered at six separate measurement periods. Multiple analyses of variance with repeated measures revealed no effect due to group assignment. Although significant changes over time were observed on most measures, these changes were statistically independent of whether the respondents were interviewed or not. The presence of interviewers was, however, associated with higher and more consistent completion rates at each measurement period.